1. Technical Field
The present invention relates to an image processing device, an image forming system, an image processing method and a computer readable medium storing a program.
2. Related Art
To perform a large volume of printing at a high speed, generally, a continuous form printer (fanfold paper printer) that prints on continuous-form print paper is used. When high-speed printing is carried out with such a continuous form printer, processing for vector to raster (bitmap) data conversion (rasterizing) in a RIP (Raster Image Processing) part is a bottleneck in processing speed. There is a need to boost the processing speed of the RIP.
Print jobs described in a Page Description Language (PDL), particularly, those for independent pages are relatively easy to divide page by page. Hence, a method in which a PDL job file is divided and then processed in parallel by plural RIP parts is proposed.
According to such a method, a print job made up of plural pages is divided page by page and processed in parallel by plural RIP parts. Thus, print job processing can be done at a higher speed in comparison with a case where rasterizing for all pages is performed by a single RIP part.
However, there is a print job in PDL for independent pages, wherein a common drawing object is used by plural pages. If a print job of this type is divided page by page, the drawing object, commonly referred to by plural pages, has to be copied to each page into which the job is divided. Thus, due to copying data such as a common drawing object to each page into which the job is divided, additional processing time is taken for dividing a job. In addition, in each RIP part, overhead for initialization or the like increases. This situation nullifies the merit of job splitting and parallel processing.
For a print job in which pages cannot be identified before receiving the whole print job, the above method also poses a negative effect that print job delivery to each RIP part is to take place only after the whole print job has been received.
For example, in PDF (Portable Document Format) jobs, information for page extraction is appended to the end of a job. For a large PDF job in which it takes some time to complete its reception, job splitting cannot begin until its transfer is completed and this results in a throughput delay.
Moreover, the above method distributes all pages of a job at the start of the job and, hence, the processing loads (load balancing) across the pages cannot be adjusted and the job processing time becomes equal to the processing time of the slowest RIP part. This method cannot apply to a flexible system in which the order of pages to print may be dynamically changed as with an in-position function.